JP3493146B2 - Image color processing apparatus, image color processing method, and recording medium - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image color processing apparatus for extracting sky color from a color image, an image color processing method, and a recording medium recording a computer program for realizing the same.

[0002]

2. Description of the Related Art Generally, in a landscape image including an image of the sky, the color of the sky greatly affects the overall atmosphere. Therefore, it may be desired to change the color of the sky to a desired color.

When a plurality of landscape images including a plurality of sky images are printed on the same page of a printed matter or on two pages of spreads, it is often desirable that the sky colors of the images are unified. .

However, the sky color of the image which is the basis of such an image may be different for the following reasons. That is, this is the case when the timing of shooting, the brand of film, and the like are different.

Therefore, in order to unify the sky color of such a plurality of images, it is necessary to correct only the sky color of the base image. Conventionally, Japanese Patent Nos. 2740436 and 2740443 have been disclosed. The operator corrects the sky color of each landscape image one by one by using a color correction method such as.

[0006]

By the way, in order to correct a color in the above method, it is generally necessary to specify a representative color and a range of colors in an image to be corrected.
Then, conventionally, the worker manually extracts these pieces of information about each landscape image one by one, so that a lot of labor cost and working time are spent.

The present invention is intended to overcome the above-mentioned problems in the prior art, and realizes an image color processing apparatus, an image color processing method, and an image color processing method capable of suppressing the cost due to labor costs and shortening the working time. It is an object of the present invention to provide a recording medium recording a computer program for.

[0008]

In order to achieve the above object, an apparatus according to claim 1 of the present invention comprises a dividing means for dividing a color image into a plurality of sections, and a representative color of each of the plurality of sections. A representative color extracting means for extracting a blue color group set for extracting a blue color group set which is a group of color blocks each having a representative color of a plurality of color blocks; A non-edge selection unit that extracts a non-edge partition set that is a set and each non-edge partition of the non-edge partition set are sequentially set as target partitions, and the spread of non-edge partitions that continuously exist around the target partition is blue. And a sky-blue determining unit that determines a representative color of the non-edge section having the largest blue degree in the set of non-edge sections as a representative sky blue of the color image.

An apparatus according to a second aspect of the present invention is the image color processing apparatus according to the first aspect, in which the representative sky blue and hue of the non-edge partition set are within a predetermined threshold. Sky blue section selecting means for selecting the sky blue section,
Lightness / darkness determining means that sets a substantially brightest representative color and a substantially darkest representative color among the representative colors of the selected sky blue section as the brightest sky color and the darkest sky color, respectively.

An apparatus according to claim 3 of the present invention is the image color processing apparatus according to claim 1 or 2, further comprising a color image based on a representative sky blue color selected. A color correction unit for correcting the color of the sky is provided.

According to a fourth aspect of the present invention, a method of dividing a color image into a plurality of sections, a representative color extracting step of extracting a representative color of each of the plurality of sections, and a plurality of sections. A blue selection step of extracting a blue section set whose representative colors are substantially blue sections, and a non-edge section extraction of a non-edge section set that is a set of non-edge sections that do not include an edge image in the blue section set. A selection step, a step of sequentially determining each non-edge section of the set of non-edge sections as a target section, and a step of determining a blue width that determines a spread of non-edge sections continuously existing around the target section as a blue width; A sky blue determination step of setting a representative color of a non-edge section having the largest blue color in the edge section set as a representative sky blue of the color image.

A fifth aspect of the present invention is the image color processing method according to the fourth aspect, wherein the representative sky blue and hue of the non-edge partition set are within a predetermined threshold. A sky-blue selection step of selecting a sky-blue section, and a light-dark selection step of selecting a substantially brightest representative color and a substantially darkest representative color of the representative colors of the selected sky-blue section as the brightest sky color and the darkest sky color, respectively. .

A method according to a sixth aspect of the present invention is the image color processing method according to the fourth or fifth aspect, further comprising a color image based on a representative sky blue color selected. A step of correcting the color of the sky is provided.

Further, a recording medium according to claim 7 of the present invention is a recording medium in which a program for extracting sky color from a color image by a computer is recorded, and the color image is divided into a plurality of sections. A division function, a representative color extraction function that extracts each representative color of a plurality of sections, a blue selection function that extracts a blue section set that is a set of sections whose representative colors of a plurality of sections are almost blue, and a blue section A non-edge selection function that extracts a non-edge partition set that is a set of non-edge partitions that does not include an edge image in the set, and each non-edge partition of the non-edge partition set as the partition of interest sequentially, while surrounding the partition of interest. A blue width determination function that finds the spread of consecutive non-edge partitions as the blue width, and a representative color of the non-edge partition with the largest blue width in the non-edge partition set It records the program for realizing the azure determination function of a representative azure color image.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

<1. System Configuration of Embodiment> FIG. 1 is a system configuration diagram of an image color processing apparatus according to an embodiment of the present invention. The image color processing apparatus will be described below with reference to FIG.

The apparatus of the present invention is an apparatus for automatically obtaining the sky color and the range of gradations of the sky color for an image.

As shown in FIG. 1, the image color processing apparatus 100
Is a CPU 1 (corresponding to a dividing unit, a representative color extracting unit, a blue selecting unit, a non-edge selecting unit, a blue color determining unit, a sky blue determining unit, a sky color section selecting unit, and a light and dark determining unit, and a color together with an input unit 5 described later. It corresponds to the correction means),
A ROM 2 in which a basic program is stored, a RAM 3 in which an image conversion program and the like described in detail below are stored, a reading unit 4 for reading out the image conversion program from a recording medium 4a such as a magnetic disk or a CD-ROM, a keyboard, a mouse, etc. Input unit 5, which receives input from the target image,
A color display 6 for displaying the image color processing result, a hard disk 7 for storing required programs and data, and the like are appropriately connected to the bus line B via an interface (I / F).
This is a device realized by the internal CPU 1 or the like executing an image conversion program in one general computer system including peripheral devices such as a scanner 8 and a printer 9 in addition to a main body part connected by L. .
It should be noted that this computer system can communicate with other devices (not shown) through the bus line BL to exchange various data and the like. Also,
The image conversion program may be stored in the ROM 2 or the hard disk 7 in advance and read out and stored in the RAM 3 for use. In that case, the ROM 2 or the hard disk 7 functions as a recording medium.

The following image color processing is performed by the above device configuration.

<2. Image Color Processing of Embodiment> Next,
Image color processing according to this embodiment will be described.
FIG. 2 is a flowchart showing the entire image color processing of this embodiment. This image color processing focuses on the blue part of the sky image included in the image, and obtains the color of such a part (a representative sky blue, the brightest sky blue and the darkest sky blue described later). Then, the blue portion of the sky is distinguished from other images by utilizing the property that the image is a smooth image that does not include an edge component whose gradation changes rapidly. The image color processing will be described below with reference to FIG. It is assumed that the operator has confirmed that the image data for which the sky color is to be extracted is prepared in advance and the sky portion is included in the image prior to the following processing. Unless otherwise specified, the CPU 1 executes the image color processing program read in the RAM 3 to automatically perform the following processing in software.

First, a color image for praying (hereinafter referred to as "target image") is divided into a plurality of sections (step S).
1). FIG. 3 is a diagram showing how the target image is divided in this embodiment. As shown in the figure, in this embodiment, the target image is divided into rectangular sections that are equal to each other. Then, two-dimensional orthogonal coordinates XY are set in the image, and each section and each pixel of these sections are designated by coordinate values. With respect to the target image, M × N sections are set by dividing M in the X direction and N in the Y direction. Hereinafter, for convenience, the X direction and Y of each section
The number of pixels in the direction is represented by a and b, respectively. In FIG. 3, M = N = 4.

By the way, although the present invention is premised on image color processing for a color image, in this embodiment, a color image is intended to have R (red), G (green) and B (blue) color components. I am trying.

Next, the representative color of each section image is obtained (step S2). Specifically, one of all the partitions is sequentially processed (hereinafter, the partition to be processed in each step is referred to as “target partition”), and three colors of each pixel of the target partition are set. Create a histogram based on the gradation of each component (has a rectangular parallelepiped class (class block) in the RGB color space because each of the R, G, and B components has a range of class values). The maximum class block, which is a class block with a large frequency, is obtained, and the average value of the gradation values for each color component of the pixels belonging to the maximum class block and the adjacent class block is used as each color component value of the representative color of the target section. is there. This representative color extraction method is disclosed in Japanese Patent Application No. 10-94569 filed by the present applicant.

Next, a set of blue sections whose representative colors in the target image are substantially blue sections is obtained (step S3).

Specifically, the representative color of each section in the target image is represented by the RGB color system as described above (representative color r
p (RGB)), this is converted into the notation (representative color rp (HSL)) in the HSL color system. And H of each section
Representative colors represented by SL system are H (hue), S (saturation), L
Whether or not each component of (brightness) is within a predetermined numerical range determines whether or not it is blue. For example, the ranges of H, S, and L components are 0 ≦ H ≦ 360 and 0 ≦ S ≦ 100, respectively.
(At L = 50), if 0 ≦ L ≦ 100,
If each component of the representative color rp (HSL) is in the range of 250 <H <290 and S> 20 and 20 <L <85, it is determined to be blue, and the partition is set to blue. It is judged that it is. The representative color rp (HSL) of each section obtained here is stored in the RAM 3.

In order to convert the values of the RGB color system into the values of the HSL color system, the values of R, G and B are converted into the three attribute values of the perceptual color in the look-up table prepared in the ROM 2. Convert to a certain hue H, saturation S, and lightness L. However, R,
The values of G, B do not directly correspond to the values of hue H, saturation S, and lightness L. Therefore, in this conversion, the values of R, G, B are once changed to the Lab color system (corresponding to the lightness L 1 The dimensional coordinates, the saturation S, and the hue H are converted into values in a three-dimensional coordinate space consisting of two-dimensional coordinates drawn on the planes of the symbols a and b. Next, the values of the Lab color system are converted into the values of the HSL color system. Figure 4
FIG. 6 is a diagram showing a relationship between a Lab color space and an HSL color space. At this time, the Lab color space and the HSL color space have a relationship as shown in FIG. In FIG. 4, the L axis of the Lab color space directly corresponds to the lightness L of the HSL color space, and a
The radius vector on the two-dimensional coordinates of the axis and the b axis corresponds to the saturation S, and the angle θ between the a axis and the radius vector corresponds to the hue H. The details of this conversion are also described in Japanese Patent Application No. 10-94569 filed by the present applicant.
Is disclosed in.

Next, for each blue section in the obtained blue section set, a set of non-edge sections that do not include an edge image and belong to a smooth portion is obtained (step S4). That is, since the sky blue image is a smooth image that does not include an edge component, a non-edge partition is obtained as a candidate for a partition including the sky blue image. Specifically, it is obtained as follows.

First, in order to determine whether the blue section belongs to a smooth portion or a portion including an edge image,
A parameter d (i, j) indicating smoothness is set for each pixel. FIG. 5 is a diagram showing a pixel designation method. As shown, each pixel is represented by P (i, j), and the indices of the X and Y components are represented by i and j, respectively. And
The parameter d (i, j) is associated with the pixel P (i, j). Here, the parameter d (i, j) is an amount representing the smoothness of the image in the pixel P (i, j), and is determined by the following conditional expression.

[0029]

[Equation 1]

[0030]

[Equation 2]

[0031]

[Equation 3]

[0032]

[Equation 4]

Note that here, R, G, and P of the pixel P (i, j)
B component is Pr (i, j), Pg (i, j), Pb (i,
j). As can be seen from these equations, the parameter d (i, j) is the pixel P (i-1, j-1), P (i + 1, j-1) that is diagonally adjacent to the pixel of interest P (i, j). ) 、
It is the absolute value of the difference between the average value of the gradations of P (i-1, j + 1) and P (i + 1, j + 1) and the gradation of the target pixel P (i, j). Therefore, the closer the value of the parameter d (i, j) is to “0”, the smaller the gradation difference between the pixel of interest and the surrounding pixels, which means that the image is smooth and does not include edge components. To do. Then, among the pixels included in the target section, such a parameter d (i, j) is
It is calculated for all pixels except the pixels in the peripheral portion of the target section.
Except for the peripheral pixels, the number of adjacent pixels to the peripheral pixels is small, and it is inappropriate to include them in the smoothness determination target. This is because even if it is excluded, it does not significantly affect the determination result.
Note that the parameter d (i, j) does not take a negative value because it is the sum of absolute values.

FIG. 6 shows a parameter d (i,
It is a flow chart which shows the procedure which calculates j). Less than,
A specific method of obtaining the parameter d (i, j) will be described.

First, the parameter d (i, j) of each pixel P (i, j) other than the periphery (outermost) included in the target section is obtained (FIG. 6: step S21). That is, 0 <i <(a
-1), 0 <j <(b-1) over the range of the indexes i and j for each pixel of interest P (i, j) based on the above equations 1 to 4, the parameter d (i, j j) is obtained and stored in the RAM 3.

Next, the sum sd of the parameters d (i, j) of those pixels P (i, j) is obtained (FIG. 6: step S2).
2). That is, the sum of the parameters d (i, j) is calculated,
Let that value be the sum sd.

Next, all the pixels P (i,
The average value of the parameter d (i, j) of j) is calculated (FIG. 6: step S23). That is, the sum sd is divided by the number (a−2) × (b−2) of all pixels except the peripheral pixels of the target section to obtain the average value ad of the parameters d (i, j).

Next, it is determined whether or not all the partitions have ended (FIG. 6: step S24), and while changing the target partition, the processing of steps S21 to S24 is performed until the average value ad is found for all the partitions (M × N). repeat. The sum sd and the average value ad are values representing the smoothness of each section.

Next, using the obtained average value ad, a procedure for obtaining a non-edge partition as a partition having an image that is smooth and does not include an edge component among all the partitions is described with reference to the flowchart of FIG. Explain.

First, a non-edge partition candidate whose average value ad is the first threshold value th1 or less is obtained (FIG. 7: step S31). Here, the first threshold th1 is a value that is determined in advance by experimentation or the like so that a section image having an average value ad larger than the threshold value has a clear edge component.

Next, the average value ad of all non-edge partition candidates
Is obtained (FIG. 7: Step S3)
2). Then, the discriminant analysis is performed on the obtained histogram to obtain the second threshold value th2 (FIG. 7: step S3
3). The discriminant analysis performed here will be described more specifically. Among the non-edge partition candidates obtained in step S31, the partition whose average value ad is the first threshold value th1 or less is a partition in which it is difficult to determine whether or not there is an edge component in the image. Such a set of sections includes a section including edge components and a smooth image section that does not include the edge components. Therefore, the obtained histogram is roughly divided into two groups: a section containing edge components, that is, a section having a large average value ad, and a section not containing edge components, that is, a section having a small average value ad. It is considered to be a thing.
Therefore, here, as a discriminant analysis, the histogram is divided into two groups so that the distance between the centers of gravity (the frequency is the weight) of the two groups is the maximum, and the average value ad that is the boundary between the two groups is obtained. The second threshold th2 to be obtained is used.

Next, a section whose average value ad is equal to or less than the second threshold value th2 is set as a non-edge section which is a section having a smooth image (FIG. 7: step S34). That is, of the two groups obtained as described above, the section belonging to the group with the smaller average value ad is the non-edge section.

As can be seen from the above, in this embodiment, attention is paid only to the sections where the average value ad is the first threshold value th1 or less, that is, it is difficult to determine whether or not there is an edge component in the image. The partitions are discriminated in more detail as to non-edge partitions. Therefore, the first threshold t
The non-edge partition can be extracted more accurately and efficiently than the case where the histogram of the average value ad is created for all the partitions without using h1 and the non-edge partition is obtained by performing the above-mentioned discriminant analysis on the histogram. .

Returning to the explanation of FIG. Next, for each section image included in the set of non-edge sections, the blue spread coefficient c (m,
n) is obtained (step S5).

FIG. 8 is a flow chart showing the procedure for obtaining the maximum blue color width coefficient. Hereinafter, a method for obtaining the maximum blue color width coefficient will be described in more detail with reference to FIG.

First, all partitions (M × N) in the target image
, A blue area coefficient c (m, m, which indicates the area in which a non-edge section (which is, of course, a blue section) exists around the section.
n) is assigned. Here, m and n represent coordinates of a section in the target image, and take values of m = 1 to M and n = 1 to N. This blue area coefficient c (m, n) exists around a partition that belongs to a set of non-edge partitions as a partition of interest,
Similarly, it represents the number of sections that belong to the non-edge section.
That is, it means that the larger the blue color area coefficient c (m, n) is, the more non-edge sections are in the periphery.

First, the blue area factor c (m, m,
As the initial value of n), "1" is set to the non-edge section and "0" is set to the other sections (FIG. 8: Step S4).
1).

Next, "1" is set to the parameter T for obtaining the maximum blue color area coefficient described later (FIG. 8: step S42).

Next, the section of c (m, n) = T is set as the target section (FIG. 8: step S43).

FIG. 9 is a diagram showing how the blue width coefficient c (m, n) of each section is set. Each rectangle represents a section B (reference numerals are partially omitted), and the numerical value in each rectangle is Blue area factor c
Represents (m, n). As shown in FIG. 9A, initially, the blue area coefficient c (m, n) of each section is "1" or "0".

Next, it is judged whether or not the blue spread coefficient c (m, n) is positive in a majority of the sections in the vicinity of (T + 2) × (T + 2) of the target section (FIG. 8: step S44). If this condition is satisfied, the blue area coefficient c (m, n) of the target section is set to T + 2.
(FIG. 8: Step S45), and the process proceeds to Step S46. On the contrary, if the condition is not satisfied, the step is directly performed in step S46.
Proceed to.

This determination will be described based on the specific example of FIG.
First, all of the blue color spread coefficients c (m, n) in the image are "1".
For the section of (parameter T), 3 × 3 ((T +
2) × (T + 2)) in the vicinity of the section (hatched section in FIG. 9A), the blue area coefficient c (m, n) is examined. Specifically, if the majority of the sections in the vicinity of 3 × 3 of the target section AB have a positive blue area coefficient c (m, n), the blue area coefficient c (m of the target section AB is , n) is set to “3”, and otherwise (“1” remains).

Next, it is judged whether or not the above processing is completed for all the sections of c (m, n) = T (FIG. 8: step S).
46), if not completed, the process returns to step S43, and the processes of steps S43 to S46 are repeated for all the sections of c (m, n) = T. Then, when it is determined that the processing is completed for all the partitions, it is determined whether or not there is a partition of c (m, n) = T + 2 (FIG. 8: step S4).
7). If there is such a section, the parameter T
Is incremented by 2 (FIG. 8: step S48), the process returns to step S43, and the processes of steps S43 to S48 are repeated until it is determined that there is no partition of c (m, n) = T + 2.

The description of the specific example of FIG. 9 will be continued again. As described above, step S43 is performed with the parameter T = 1.
~ The process of S46 is c (m, n) = 1 (that is, c (m, n) =
FIG. 9B shows a state in which all the sections of T) are performed.
Is. As shown in FIG. 9B, since there is a section of c (m, n) = 3 (that is, c (m, n) = T + 2), it is determined to be “Yes” in step S47, and T = T in step S48. Then, the process returns to step S43. In step S44, this time, all blue width factors c in the image are
For a section where (m, n) is “3” (T), the blue area coefficient c (m, n) of the area in the vicinity of 5 × 5 is examined, and the blue area coefficient c (m, n) is positive. If the number of sections that take a value is the majority, the blue area coefficient c (m, n) of the section is set to "5" (T + 2). This state is shown in FIG. 9 (c). In this way, the range for detecting the spread of the non-edge section (also the blue section) around each section is gradually increased (the parameter T is increased by “2”), and as described above. The value of the blue color width coefficient c (m, n) is updated. As a result, the blue width coefficient c (m, n) of the area having a large spread of the non-edge areas having a blue color is gradually increased, and the blue area coefficient c is gradually increased.
The number of sections in which the value of (m, n) is updated decreases.

Then, in the determination at step S47, "N
If it is determined to be “o”, the parameter T is set as the maximum blue color area coefficient (FIG. 8: step S49). That is, the spread of the blue non-edge partitions has been detected for all the partitions, and the blue width factor c of all the partitions has been detected.
Since the maximum value of (m, n) should be the parameter T at that time, the parameter T is set to the maximum blue color range coefficient.

Next, returning to FIG. 2, the section having the maximum blue area coefficient (hereinafter referred to as "maximum blue area") is obtained as the maximum blue section (step S6), and the representative color of the maximum blue section is calculated as follows. It is determined to be a representative sky blue that is a representative sky color of the image and stored (step S7). In other words, the blue image of the sky is considered to be the part where the blue and smooth image spreads most widely in the whole image, so it is a typical non-edge partition and the partition with the widest spread of non-edge partitions around it. The color is a typical sky blue. However, there is only one segment where the blue color factor c (m, n) matches the maximum blue color factor.
In one case, the representative color of the section is the representative sky blue of the image, but there may be a plurality of matching sections. In such a case, the representative color rp of those sections
The average value of the hue (H) of (HSL) is calculated. Then, of the representative colors rp (HSL), the representative color having the hue closest to the average value of the obtained hues is determined as the representative sky blue of the image. Then, the obtained typical sky blue is stored in RAM
Store in 3.

Now, a representative sky blue color of the target image is obtained.

Next, of the non-edge partition sets, a sky-blue partition set which is a set of typical sky-blue partitions having a hue within a predetermined threshold value is obtained (step S8). In particular,
When the difference between the hue (H) of the representative color rp (HSL) and the hue (H) of the representative sky blue RP (HSL) in the set of non-edge sections is less than or equal to a predetermined threshold value, the representative color is the representative sky blue. Assuming that the colors are close to each other, the set of the sections is obtained as the sky blue section set.

Next, among the representative colors of the sky-blue sections of the obtained sky-blue section set, the brightest representative color and the darkest representative color are respectively the brightest sky color and the most bright sky color which are the lightest sky colors of the target image. The darkest sky color which is the color of the dark sky is determined and stored (step S9). Specifically, the representative color of the brightest (L is maximum) section that belongs to the sky-blue section set is the brightest sky-blue. Similarly, the representative color of the darkest (L is the smallest) section that belongs to the sky-blue section set is the darkest sky-blue. Then, the brightest sky blue and the darkest sky blue are RA.
Store in M3.

As described above, the representative sky blue, the brightest sky blue and the darkest sky blue were obtained. Then, if necessary, the sky color is corrected using these color data (step S10).

Specifically, the representative sky blue is used as the color to be corrected, and the brightest sky blue and the darkest sky are used as the range of color correction processing with the color to be corrected as a reference. Convert to color system. In addition, the worker
The target color, which is the corrected color represented by the L color system, is designated through the input unit 5. Next, parameters for color correction are set from the representative sky blue and the target color. This parameter is a value such as the ratio of a typical sky blue and the target color. Further, the pixel-by-pixel image data of the target image is converted into HSL color system image data in the same manner as described above, and it is determined whether or not the converted image data falls within the range of the color correction processing. Only for the image data to be processed, the original RGB image data of the pixel is multiplied by the above parameters to correct the target color or a color close to it. Details of this color correction method are described in Japanese Patent No. 274 by the present applicant.
It is disclosed in Japanese Patent No. 0436 and Japanese Patent No. 2740443.

It is also possible to record the color data of the representative sky blue, the brightest sky blue and the darkest sky blue on the hard disk 7, the recording medium 4a or the like and retrieve them when necessary for color correction.

In some cases, a plurality of target images may be prepared in the above, and in that case, the above processing is repeated. This is the end of the image color processing according to the embodiment.

As described above, according to the image color processing apparatus of this embodiment, the typical sky blue, the brightest sky blue and the darkest sky blue are automatically obtained for the color image, and based on them, Since the sky color of the color image is corrected, it is possible to reduce the work load as compared with the case where the sky color of the color image is manually extracted and the color is corrected, thereby reducing the labor cost. It is possible to reduce the work time.

Further, since the representative color of the non-edge section having the largest blue color in the set of non-edge sections is set as the representative sky blue of the color image, even if the sky color has shades, the representative color is selected. The sky blue can be extracted accurately. Further, by performing color correction of the sky image based on such a representative sky blue, color correction can be performed based on a color suitable for the sky image, and the quality of color correction can be improved.

Further, when obtaining the blue color spread coefficient, it is determined whether or not each of the H, S, and L components of the representative color rp (HSL) of each section is within a predetermined range. Therefore, it is possible to objectively and accurately capture the blue color of the sky and its spread.

Further, according to this embodiment, a sky blue partition having a representative sky blue and a hue within a predetermined threshold value is selected from the non-edge partition set, and the brightest representative color and the most representative color of the sky blue partitions are selected. Since the representative dark colors are the brightest sky blue and the darkest sky blue, respectively, the brightest sky blue and the darkest sky blue can be used to accurately specify the sky color range. It is possible to easily and accurately extract and correct only the colors of the image. Further, thereby, the cost due to the labor cost can be further suppressed, and the working time can be further shortened.

<3. Modifications> In the above embodiment, an example of the image color processing apparatus and the image color processing by the image color processing apparatus is shown, but the present invention is not limited to this.

For example, in the above embodiment, the image color processing is performed using the image data that is the target of sky color extraction as it is, but the image color processing is started (step S1 in FIG. 2 is executed). The image color processing may be performed after comparing the size of the target image data with a predetermined threshold value, and if the size is larger than the predetermined threshold value, the size is reduced by an appropriate method (pixel thinning, etc.). In that case, the data amount of the image data is reduced, so that the storage capacity of the image data such as the RAM is reduced and the processing speed is also improved.

Further, in the above embodiment, as a method of obtaining the parameter d (i, j) representing the smoothness of the image of each pixel, the pixel of interest P
Pixel P (i−1, j−) that is diagonally adjacent to (i, j)
1), P (i + 1, j-1), P (i-1, j + 1), P (i +
The absolute value of the difference between the average value of the gradations of (1, j + 1) and the gradation of the pixel of interest P (i, j) is obtained, and the sum thereof is used as the parameter d (i, j) by the equation (3). However, the target pixel P
Pixels P (i−1, j), P (i + 1, j), P (i, j−) that are adjacent to the positive and negative sides of the X axis and the Y axis with respect to (i, j).
1), the average value of the gradations of P (i, j + 1) and the target pixel P (i,
j) the absolute value of the difference from the gradation is used as the parameter d (i, j), or the pixel P (i + r, j + s) adjacent to the pixel P (i, j) is used.
(R = 0, ± 1 and s = 0, ± 1) The absolute value of the difference from the average value of all gradations is used as a parameter d (i, j), and further, in addition to the most adjacent pixel, You may use the absolute value of the difference with the average value of the gradation of the pixel adjacent to the outer side.

Further, in the above-described embodiment, the representative colors of the brightest section and the darkest section are respectively the brightest sky blue and the darkest sky blue in step S9 of the flowchart of FIG. A light color and a color slightly darker than the representative color of the darkest section may be the brightest sky blue and the darkest sky blue, respectively. This is shown in Figure 2.
Since the representative color of each section image extracted in step S2 in step S2 extracts the representative color of those sections, not all the pixels in those sections have the representative color, and normally, It seems to have a spread in a predetermined range around the representative color. Therefore, it is considered that some brighter (or darker) pixels are also present in the partition having the brightest (or darkest) representative color. At the time of correction, even those pixels that are originally included in the background image in those sections are not color-corrected (remaining uncorrected), and conversely, pixels that are not originally included in the sky color are also color-corrected. This is because things (overcorrection) may occur.

[0072]

As described above, according to the first to third aspects of the present invention, a dividing means for dividing a color image into a plurality of sections and a representative color for extracting a representative color of each of the plurality of sections. Extraction means, blue selection means for extracting a blue section set that is a set of sections whose representative colors are substantially blue, and non-edge that is a set of non-edge sections that do not include an edge image in the blue section set A non-edge selection unit that extracts a partition set and each non-edge partition of the non-edge partition set are sequentially set as a target partition, and a spread of non-edge partitions that continuously exist around the target partition is obtained as a blue width. Since the area determining means and the sky-blue determining means for making the representative color of the non-edge section having the largest blue degree of the non-edge section set the representative sky blue of the color image, the color image is manually It is possible to reduce the burden work as compared with the case to extract the color of the sky, it makes it possible to reduce the cost of labor costs, can shorten the working time.

Further, according to the inventions of claims 1 to 7, since the representative color of the non-edge section having the largest blue width in the set of non-edge sections is the representative sky blue of the color image, Even if there are shades of color, a typical sky color can be accurately extracted from them. Further, by correcting the sky color based on such a representative sky blue, the color correction can be performed based on a color suitable for the sky color, and the quality of the color correction can be improved.

Further, according to the inventions of claim 2, claim 3, claim 5, and claim 6, a representative sky blue color and a sky color color zone having a hue within a predetermined threshold value are selected from the non-edge partition set, and they are selected. The brightest and darkest representative colors of the representative colors in the sky blue section are the brightest sky color and the darkest sky color, respectively, so the sky color range is accurately specified using the brightest sky color and the darkest sky color. Therefore, even when the sky color is corrected, only the color of the sky image can be easily extracted and corrected. Further, thereby, the cost due to the labor cost can be further suppressed, and the working time can be further shortened.

Further, according to the invention of claim 3, since the color correction means for correcting the sky color of the color image based on the selected representative sky blue is provided, the sky color is automatically corrected. Since it can be performed, the labor cost can be further suppressed and the working time can be shortened.

[Brief description of drawings]

FIG. 1 is a system configuration diagram of an image color processing apparatus that is an embodiment of the present invention.

FIG. 2 is a flowchart showing the entire image color processing of the embodiment.

FIG. 3 is a diagram showing how a target image is divided in the embodiment.

FIG. 4 is a diagram showing a relationship between a Lab color space and an HSL color space.

FIG. 5 is a diagram showing a method of specifying pixels.

FIG. 6 is a parameter d of a target section according to the embodiment.
It is a flowchart which shows the procedure which calculates | requires (i, j).

FIG. 7 is a flowchart showing a procedure for obtaining a non-edge section in the embodiment.

FIG. 8 is a flowchart showing a procedure for obtaining a maximum blue color width coefficient in the embodiment.

FIG. 9 is a diagram showing how to set a blue width coefficient of each section according to the embodiment.

[Explanation of symbols]

1 CPU 2 ROM (recording medium) 3 RAM 4 reading section 4a recording medium 7 Hard disk (recording medium) 100 image color processing device Section B AB attention section T parameter c (m, n) Blue area factor th1, th2 first and second thresholds

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06T 1/00 510 G06T 11/60 120 G06T 1/00 280 G06T 3/00 400 H04N 1/46 H04N 1 / 60

Claims (7)

(57) [Claims] 1. A dividing unit that divides a color image into a plurality of sections, a representative color extracting section that extracts a representative color of each of the plurality of sections, and a representative color of each of the plurality of sections that is substantially blue. A blue selection unit that extracts a blue partition set that is a set; a non-edge selection unit that extracts a non-edge partition set that is a set of non-edge partitions that does not include an edge image in the blue partition set; Of each of the non-edge partitions, the blue width determining means for determining the spread of the non-edge partitions that continuously exist around the target partition as the blue width, An image color processing apparatus comprising: a sky blue determination unit that sets a representative color of the non-edge section having the largest blue color to a representative sky blue of the color image. 2. The image color processing device according to claim 1, further comprising a sky blue section selecting unit that selects a sky blue section whose hue and hue are within a predetermined threshold value from the non-edge section set. An image color processing device comprising: a bright and dark representative color that is a brightest representative color and a darkest representative color of the representative colors of the selected sky blue section, respectively. . 3. The image color processing apparatus according to claim 1, further comprising color correction means for correcting the sky color of the color image based on the selected representative sky blue. An image color processing device comprising: 4. A division step of dividing a color image into a plurality of sections, a representative color extraction step of extracting a representative color of each of the plurality of sections, and a section of a section in which each of the plurality of sections has a substantially blue representative color. A blue selection step of extracting a blue partition set that is a set; a non-edge selection step of extracting a non-edge partition set that is a set of non-edge partitions that does not include an edge image in the blue partition set; Of each of the non-edge partitions, the blue width determining step of determining the spread of the non-edge partitions continuously existing around the target partition as the blue width, and the non-edge partition set among the non-edge partition sets. A sky color determination step of setting a representative color of the non-edge section having the largest blue color as a representative sky blue of the color image. 5. The image color processing method according to claim 4, further comprising a sky blue selection step of selecting a sky blue section having a hue within a predetermined threshold value of the representative sky blue from the non-edge section set. An image color processing method, comprising: a bright and dark selection step of selecting a substantially brightest representative color and a substantially darkest representative color of the representative colors of the selected sky blue section as the brightest sky color and the darkest sky color, respectively. . 6. The image color processing method according to claim 4, further comprising a step of correcting the sky color of the color image based on the selected representative sky blue. An image color processing method characterized by: 7. A recording medium recording a program for extracting sky colors from a color image by a computer, a dividing function for dividing a color image into a plurality of sections, and a representative color of each of the plurality of sections. A representative color extraction function, a blue selection function for extracting a blue section set that is a set of sections whose representative colors of the plurality of sections are substantially blue, and a non-edge section that does not include an edge image in the blue section set. A non-edge selection function that extracts a set of non-edge partitions, and a non-edge partition that continuously exists around the target partition while sequentially setting each non-edge partition of the non-edge partition set as a target partition And a blue color determination function that obtains the blue color as the blue color, and the representative color of the non-edge partition having the largest blue color of the non-edge partition set is the color image. A computer-readable recording medium in which a program for realizing a sky-blue determination function that is a typical sky-blue of an image is recorded.